Re: [Trans] Starting to think about IETF 90

[Dmitry Belyavsky <[e-mail]>]

Hello Melinda,


On Tue, Jun 3, 2014 at 12:47 AM, Melinda Shore <[e-mail]>
wrote:

>
> So, if you've got a document that you're trying to progress,
> now's a very good time to post to the mailing list with a
> pointer to the document, a quick summary of what you're
> trying to do, and what you think the open issues are or
> areas that need discussion.
>

Unfortunately, I will not be present in Toronto.
So I think there are the following problems that are the subject to solve
for putting the CT into real world.

1. The behaviour of CT-supporting clients is underspecified. There are a
lot of corner cases that should be a subject for clarification. I have sent
the draft of my vision mixed with mostly Ben’s answers about the client
behaviour to the Trans list some time ago. As there is no objections or
agreements, I think that my suggestions should be either ignored and
replaced by a specification written from scratch or adopted as a part of
RFC after discussion.
Also the absence of the gossip protocol description disturbs me too.

2. There is a lack of understanding of infrastructure which is necessary
for the deployment of CT. If we get a hundred of logs, it is not better
than a hundred of CAs we have now. If we found a misbehaviour of log or of
a CA, it should be reported to - where? As far as I understand, this is out
of scope for IETF, but who will establish the rules here?

The adoption of the CT to operate together with DNSSec seems to be a
separate topic and I think, there should be a separate document covering it.


Thank you!
-- 
SY, Dmitry Belyavsky